How Artificial Intelligence Can Help Detect Wildfires
“Forests are the lungs of our land, purifying the air and giving fresh strength to our people.” ~ Franklin D. Roosevelt
Over years, decades and centuries, the earth’s forests have faced many wildfires. For the period 2012-2021, there was an average of 61,289 wildfires annually and an average of 7.4 million acres impacted annually, according to statistics by the Congressional Research Service. In 2021, there were 58,968 wildfires that burned 7.1 million acres. In an era of advanced technologies such as artificial intelligence (AI), big data, analytics and cloud computing, is there a way to combat wildfires? Yes, here’s how AI can help.
What is a Wildfire?
“A wildfire is an unplanned fire that burns a natural area such as a forest, grassland, or prairie,” as defined by the World Health Organization. Wildfires are often caused by human activity or a natural phenomenon such as lightning, and they can happen at any time or anywhere. It is estimated that in 50% of wildfires recorded, it is not known how the fires started. Wildfires impact the whole ecosystem, including fauna, flora, soil, water and air.
Deaths & Economic Impact Caused By Wildfires
According to a research study, more than 33,500 deaths were attributable to wildfire pollution for the time period of January 2000 through December 2016. The repercussions take the form of huge economic losses too. The California wildfires in 2018 alone caused an economic loss of nearly $150 billion.
Data by the National Interagency Coordination Center (NICC) indicates that the number of acres affected by wildfires annually has increased over the last 30 years. According to a report, “Of the 1.5 million wildfires that have occurred since 2000, as many as 237 exceeded 100,000 acres burned and 15 exceeded 500,000 acres burned.” However, the percentage of conflagrations has not been high. The risk of wildfires increases in extremely dry conditions, such as drought and during high winds. With global warming and changing climatic conditions, the fire seasons are starting earlier and ending later each year. A study finds that “the amount of Western U.S. land burned by 'high-severity' wildfires (fires that destroy more than 95% of trees) has increased 800% since 1985.”
How AI Is Used in Firefighting
Although it is impossible to eradicate wildfires completely, better prediction, early detection and quick response can contain the extent of damages. Here’s some work being done in this field.
ALERTWildfire, a consortium of The University of Nevada, Reno, University of California San Diego, and the University of Oregon, is providing fire cameras and tools to help firefighters and first responders to a) discover, locate and confirm fire ignition, b) quickly scale fire resources up or down, c) monitor fire behavior during containment, d) help evacuations through enhanced situational awareness, and e) observe contained fires for flare-ups. Since the installation of its first alert camera in 2013, its network has grown to more than 1,000 cameras with many of them now embedded with AI as an additional analytical tool. AI-powered cameras provide invaluable inputs to differentiate between fog and wildfire smoke, especially during hot, dry and windy weather. AI can scan images and detect difference in images at a rate impossible to perceive by the human eye.
Another AI-based wildfire detection solution, FireScout has been developed by Alchera, a leading AI technology provider. FireScout enables smoke detection at early stages through a network of fire watch cameras that work round the clock. In November 2021, Pacific Gas and Electric Company (PG&E) (PCG) collaborated with ALERTWildfire to install 138 new HD cameras across high fire-threat districts, out of which 46 cameras are included in the new AI testing program in partnership with Alchera and ALERTWildfire.
Back in 2019, IBM teamed up with Compta Emerging Business to help create a fire detection product. Dubbed as Bee2FireDetection, the solution aims to better predict wildfires using AI from IBM Watson, IoT data and The Weather Company data. In November 2021, the Bee2FireDetection technology was selected by the rePLANT project for protecting the Portuguese forest by spotting forest fires, using AI.
In Brazil, Sintecsys, a commercial agriculture technology, and Omdena, a platform for building AI solutions to real-world problems, successfully built a system that identifies smoke and flames with more than 95% accuracy, which dramatically reduced false positives and the time until firefighting assistance is called onto the scene. Sintecsys and Omdena are working on the second project that will enable detection of smoke and fire outbreak in nighttime images. Sintecsys monitors 8.7 million acres across four Brazilian biomes, including amazon forest.
In November, 2021, NVIDIA (NVDA) and Lockheed Martin (LMT) joined hands with the U.S. Department of Agriculture Forest Service and Colorado Division of Fire Prevention & Control (DFPC) to fight wildfires using AI and digital-twin simulation. The two companies are building the world’s first AI-centric lab dedicated to predicting and responding to wildfires. “The lab will use NVIDIA AI infrastructure and the NVIDIA Omniverse advanced visualization and virtual world simulation platform to process a fire’s magnitude and forecast its progress. By recreating the fire in a physically accurate digital twin, the system will be able to suggest actions to best suppress the blaze,” according to the NVIDIA blog.
AI has made the processing of data (collected through sensors, cameras, satellite images, drones, among other sources) faster, cost-effective and accurate. The speed and accuracy of outcomes can provide the crucial ‘time edge’ that can go a long way in limiting wildfires and protecting our forests and its ecosystem.
Disclaimer: The author has no position in any stocks mentioned. Investors should consider the above information not as a de facto recommendation, but as an idea for further consideration. The report has been carefully prepared, and any exclusions or errors in it are totally unintentional. The data is based on multiple reports (links provided) and information based on company websites.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.